INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

estones

-0.79

 Pillar

-0.71

calling

-0.66

ples

-0.64

bery

-0.63

PE

-0.61

Bah

-0.61

 podium

-0.61

inx

-0.61

 Phill

-0.61

POSITIVE LOGITS

ingen

0.84

ortium

0.82

uggle

0.74

orah

0.74

avery

0.72

doi

0.72

 manpower

0.68

\\\\\\\\

0.68

ukong

0.68

$$$$

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.